Partial Wave
Partial Wave
Dipankar Chakrabarti
Department of Physics, Indian Institute of Technology Kanpur, Kanpur 208016, India
(Dated: April 5, 2020)
1
I. INTRODUCTION
The asymptotic form of the wavefunction determines the scattering cross-section but can-
not itself be found without solving the wave equation. As we have seen before in bound
state problem, this can be done when the wave equation is separable in different variables/co-
ordinates, e.g., spherically symmetric potential(say hydrogen atom problem). For central po-
tential V (r, θ, φ) = V (r) only and the orbital angular momentum L2 of the particle is a constant
of motion, [H, L2 ] = 0. We can construct simultaneous eigen states of H, L2 and Lz : φk,l,m (~r).
For a free particle the Hamiltonian is just the kinetic energy term, and the simultaneous eigen
function φ0klm satisfies
~2 k 2 0
H0 φ0k,l,m = φk,l,m , L2 φ0k,l,m = l(l + 1)~2 φ0k,l,m , Lz φk,l,m = m~ φ0k,l,m
2m
We often ask the question how states with definite angular momentum are affected by the scat-
terer which lead to the method of partial waves. The spherical waves with different values of
orbital angular momentum l (e.g., φ0k.l.m (~r)) are called the partial waves and the corresponding
states | k, l, mi are called the partial-wave states.
The angular dependence of wave functions(free or in central potential V (r)) is given by
Ylm (θ, φ). The potential V (r) only influences the radial dependence of the wave function.
In scatterings, we always make measurements far away from the scattering region. So, let
us consider the wave function at large r(r → ∞):
In free space, we denote the partial wave by φ0klm . The angular part of the free spherical
waves with well defined angular momentum is given by Ylm (θ, φ) and the radial part of the
eikr e−ikr
wavefunction is the superposition of r
and r
(outgoing and incoming waves) with a well
defined phase difference.
In central potential V (r), the partial wave is written as φklm . The angular part of spherical
wave is Ylm and the radial part is also superposition of the outgoing and incoming waves.
However the phase difference between these two waves is different from that for free spherical
waves(φ0klm ).
⇒ The potential V (r) introduces a phase shift (δl ). This phase difference is the only difference
between the asymptotic behavior of φklm and that of φ0klm .
2
• For a fixed k, we need to know the phase shift δl for all values of l to calculate the scattering
cross section.
• In our discussion of scattering we assumed the incident wave is a plane wave. If we want
to ask how a state with definite angular momentum is affected by the scatterer, we should be
able to write the incident wave as a superposition of orbital angular momentum states. How
to write the plane incident wave in the basis of angular momentum eigen functions?
The incident plane wave can be written as
Since there is no φ dependence in the wavefunction, only m = 0 appears in the above equa-
tion (and writing r dependence as jl (kr) is a well educated guess at this moment with the
requirement of finiteness of the wavefunction at the origin(r = 0) incorporated). To evaluate
the coefficient Cl :
Z
Cl jl (kr) = dΩYl0∗ (θ)eikr cos θ
Z
1 L−
∗
= dΩ q ( )l Yll (θ, φ) eikr cos θ
(2l)! ~
Z
1 L+
= dΩ q Yll∗ (θ, φ)( )l eikr cos θ
(2l)! ~
s
2l l! 4π Z
= q (ikr)l dΩ|Yll (θ, φ)|2 eikr cos θ (2)
(2l)! (2l + 1)!
where L± are the ladder operators. This equation is true for any r. Take r → 0 and replace
(kr)l
q
jl (kr) = (2l+1)!!
in l.h.s and eikr cos θ ∼ 1 in r.h.s, to finally obtain Cl = il 4π(2l + 1). So, we get
∞ q
eikz = il 4π(2l + 1)jl (kr)Yl0 (θ)
X
l=0
∞
il (2l + 1)jl (kr)Pl (cos θ)
X
= (3)
l=0
⇒ a state with well defined linear momentum involves all possible orbital angular momenta.
3
II. PARTIAL WAVES
p2
H0 = (4)
2µ
0
φ0klm = Rkl (r)Ylm (θ, φ). (5)
d2 2 d l(l + 1)
2
+ + (1 − ) Rkl (ρ) = 0 (7)
dρ ρ dρ ρ2
which is the spherical Bessel equation. The general solution of Eq.7 is given by
ρ→0 ρl
jl (ρ) → (9)
(2l + 1)!!
ρ→0 (2l + 1)!!
nl (ρ) → (10)
ρl+1
4
So, we can write the complete solution as
s
2k 2
φ0klm (r) = jl (kr)Ylm (θ, φ) (12)
π
ρ→∞ 1 lπ
jl (ρ) ∼ sin(ρ − ) (14)
ρ 2
ukl (r)
φklm = Rkl (r)Ylm (θ, φ) = Ylm (θ, φ) (16)
r
Finiteness of the wavefunction Rkl at the origin implies ukl = 0 at r = 0. We can consider the
eqn.(17), as an 1-D problem with effective potential
l(l + 1)~2
Vef f = V (r) + , f or r > 0, and Vef f = ∞, f or r < 0. (18)
2µr2
For scattering problems, we consider the potential V (r) to be of finite range, so for large r (i.e.,
l(l+1)
at r → ∞) V (r) → 0 and r2
→ 0 and the radial wave equation reduces to
d2
2
ukl + k 2 ukl = 0 (19)
dr
5
The solution of the above eqn is
r→∞
ukl (r) = |A|[eikr − e−ikr e2iβl ]
= Cl sin(kr − βl ) (21)
lπ lπ
Note for V = 0 we got βl = 2
. Writing βl = 2
− δl we get,
r→∞ lπ
ukl = Cl sin(kr − + δl ) (22)
2
δl is called the phase shift of the partial wave. In general, δl depends on the energy k.
So, now we have the asymptotic form of the partial wave inpresence of central potential as
We know that if we assume the incident wave is propagating along z axis, there is a symmetry
in φ (symmetry of rotation about z axis) as the potential is independent of (θ, φ). hence the
6
scaterring cross-section σ is independent of φ. Following Eq.(3), we can write the wave reaching
the detector after scattering
∞
hx | ψ + i = ψ(x) = il (2l + 1)Rkl (kr)Pl (cos θ)
X
(25)
l=0
~2 d2 l(l + 1) ~2 k 2
[− r + V (r) + ]Rkl = Rkl (r) (26)
2µr dr2 2µr2 2µ
We know that for r → ∞, the solution of Rkl (r) is given by (see Eq(22))
so,
∞
sin(kr − lπ/2 + δl )
il (2l + 1)Al
X
ψ(x) = Pl (cos θ) (28)
l=0 kr
+ ikz eikr
hx|ψ i = e + f (θ) (29)
r
∞
eikr e−i(lπ/2−δl ) − e−ikr ei(lπ/2−δl )
il (2l + 1)Al
X
⇒ Pl (cos θ) =
l=0 2ikr
∞
eikr e−ilπ/2 − e−ikr eilπ/2 eikr
il (2l + 1)
X
Pl (cos θ) + f (θ) (31)
l=0 2ikr r
Al = eiδl . (33)
7
Putting Eq(33) in Eq.(32), finally we get the scattering amplitude,
∞
1 X
f (θ) = (2l + 1)(e2iδl − 1)Pl (cos θ) (34)
2ik l=0
∞
1X
= (2l + 1)eiδl sin δl Pl (cos θ) (35)
k l=0
The derivation of f (θ) rests on the principles of (i) rotational invariance (no φ dependence)
and (ii) probability conservation( which gives the expression in Eq.(21)).
The differential scattering cross-section
dσ
= |f (θ)|2 (36)
dθ
and total scattering cross-section
Z
dσ
σ= dΩ. (37)
dθ
We can calculate the total cross-section by integrating the differential cross-section. But a
simpler way is to use the optical theorem. According to the optical theorem
4π
σ = Im[f (θ = 0)]
k
∞
4π 1X
= Im (2l + 1)eiδl sin δl Pl (cos θ)
k k l=0
∞
4π X
= (2l + 1) sin2 δl (38)
k 2 l=0
Note, σ = 0 if δl = 0 or π.
So, in partial wave analysis we need to calculate the phase shift δl for a given potential. If
we assume the range of the potential is R, i.e, for r > R the potential V (r > R) = 0, we need
to solve the Schrodinger equation where the potential V (r) 6= 0 and match that wave function
with the solution outside that region by imposing the boundary conditions at r = R.
• Read the sections "Determination of phase shifts" and "hard sphere scattering" from
Sakurai.